Friday, January 18, 2008

Oracle's BEA acquisition: SOA perspective

The bigest Software Vendors are continuosly bying other companies. The most recent acquisitions are BEA's acquisition by Oracle and the anouncement that Sun Microsystems is going to aquire MySQL.
As SOA is the main topic of this blog, I will focus on the first acquisition, which is related to SOA. MySQL acquisition process by Sun is not directly related to SOA.
BEA is a long time target for acquisition: a technology leader with significant market share which is also significantly smaller than its competitors like IBM, Microsoft or Oracle.

On January 18th 2004 I published an article in Hebrew titled: Why BEA will not be aquired by Microsoft?
This article was a response to a Research Note by Yankee Group's analyst Dana Gardner, which correctly described the acquisitions trend. However, I disagreed with the scenario of Microsoft bying BEA.
My argument was that I am not sure if any of the bigger companies will aquire BEA, but if any company will aquire BEA, it will be Oracle and not Microsoft.
BEA's main assets in 2004 were WebLogic Application Server and Tuxedo Transactions Monitor
The main reasons for thinking of Oracle as the leading candidate for aquiring BEA were:

  1. Both are J2EE market leaders (Oracle lagging behind IBM and BEA) so the acquisition could increase Oracle's installed base in comparison to the leading competitor in that market: IBM.
  2. BEA middleware products portfollio has technological advantages over similar Oracle products. One main advantage is ease of use. Some people called BEA the "Microsoft of the Java World" due to its products user friendliness and ease of development and deployment.
  3. Oracle already aquired a Swedish product which became its Oracle9IAS: Oracle's Application Server
On 2004 I thought that Microsoft could find more appropriate acquisition targets addressing its relative weaknesses:
  1. Microsoft's targets would be application vendors (ERP, CRM etc.) rather than infrastructure vendors.
  2. Microsoft will not buy a plarform vendor and especially a Java vendor for converting the installed base to DotNet.
    Probably, the installed base will continue to use the Java platform of another Java vendor, mainly IBM and Oracle, its competitors.
  3. It is too expensive ticket for influencing the Java Community Process (JCP).
  4. The weak chain in Microsoft's infrastuctue of 2004 was Integration. If the solution is an aquistion, why buying a platform vendor and not a pure integration player?
Finally, in January 2008 Oracle aquired BEA after aquiring application vendors (PeopleSoft, Siebel etc.) and Business Performance appication vendors (Hyperion) . It is part of a process of market consolidation and emergence of four leading SOA Eco Systems: Microsoft, IBM, Oracle and SAP.

Oracle's SOA challanges
Both Oracle and BEA developed SOA infrastucture architecture based on products build by the company, partnerships and acquisitions. The challnge facing Oracle is defining a comprehensive SOA architecture based on two different product portfolios and approaches.
Although SOA is designed for heterogenity Oracle will have to decide which approach will lead, which products will compose its mainstream SOA refernce model.
Let us look at some of the challenges facing Oracle:

Merging two Platforms
It is the first time oracle is buying a platform vendor offering SOA solutions including Application Server, Intgeration Solutions (AquaLogic products line including ESB, BPM etc.), Portal (PlumTree aquired by BEA few years ago), Repository (former Flashline aquired by BEA)). Oracle acquisitions experience is mainly Applications acquisitions (e.g. PeopleSoft, Siebel). The only infratsructure acquisition with some similarities to BEA acquisition is the acquisition of the databases products from Digital (mainly the RDB database) in the 1990s. However, you can not compare acquisition of a product line completing its falgship Oracle database for certain niches and adding installed base to that flagship product, to a platform vendor acquisition.

Deciding on the leading products and platform
Oracle surely will continue to market both platforms and think of merging them for the Long Run. Porbably it will have to decide which product is the leading product. For example; will it be WebLogic with larger market share or Oracle Application Server which currently lags after WebLogiic in market share?
Is it Oracle Portal or BEA Portal? Is it FusionMiddleware or AquaLogic?
If the decision should be mixed (i.e. some products from BEA SOA solution and other products composing parts of Oracle solution) another challenge will be building cohorent architecture.

Maintaining BEA's Strengths
As a smaller vendor than its competitors BEA had to find its technological advantages otherwise organizations would prefer a stronger and more stable vendor's solutions.
These advantages could be summarized in three bullets:

  • Innovation
    Original and Innovative concepts helped BEA to get advantage over competitors.
    For example BEA was the first vendor implementing the concept of APS (Application Platform Suite).
  • Standards Compliance
    BEA products comply to standards such as J2EE standards, with less proprietry extensions than its competitors. BEA also supports new standards version quicker. The result is functional rich and portable products.
  • Ease of use
    BEA's solutions ease of use and easy implementation attracts customers.
Will the 80 pounds gorilla (Oracle) be able to keep the Bee (BEA)'s Agility?

Merging the organizational Culture and preserving key experts
The Organizational culture of the company will be Oracle's and not any of the aquired apploications companies or BEA. BEA's key professionals should adapt to Oracle's culture or find a new job. The challenge is to keep most of them as Long term Oracle's employees.

First Take
I may be wrong, but my First Take predicitions are:

  1. WebLogic will be Oracle's leading Application Server competing with IBM's WebSphere application Server (It is not an installed base only acquisition)
    Due to WebLogic market share and technical features it will serve as the Services and Components execution environment and not Oracle Application Server.
    This prediction is based upon similar CRM trend after aquisition: Siebel CRM which has higher market share than Oracle CRM, broader functionality and technological strenghts is Oracle's preffered CRM solution.
    It should be noted that PeopleSoft's applications before its aquisition by Oracle were deployed on three platforms: IBM WebSphere, BEA WebLogic and Microsoft. The most frequent platform for these deployments is BEA's WebLogic.
  2. BEA's SOA solutions will dominate. Some Fusion Middleware components will be the exception in this respectSoftware AG's WebMethods aquisition is a similar example. The brandname of its SOA solutions was changed to WebMethods SOA product line, while Software AG leading SOA product the Centrasite Registry remains as a core SOA product.
  3. Increased cooperation between Oraqcle and HP and less Oracle-Sun cooperation
    HP is BEAs partner and Oracle partner which compelements BEA and Oracle's solutions. Its Mercury aquired Systinet Registry is licensed both by Oracle and by BEA.
    On the other hand, Sun's aquisition process of MYSQL database will position it as a competitor to Oracle's flagship database and less as the old times traditional partner.
  4. New Telecom Oracle applications
    The traditional ERP market is mature. Oracle will try to expand its SOA based future application generation to new markets. The Telecom sector vertical applications will be one of Oracle's new application offering
    BEA's specialized Telco platform will help Oracle in leveraging these future applications.


Dana Gardner said...

Avi -- Great points! I have to say I had very little confidence that Microsoft ever would acquire BEA, my point back in 2004 was more to show how Microaoft could become a full-service provider to heterogenous enterprises, and that BEA would allow it to do so. The fact that Microsoft's strategy has not shifted shows that SOA forms a superset to Microsoft's subset of capabilities.

Ronen said...

Hi Avi,

Good post. I have some comments which express my opinion only without having any insiders knowledge and do not necessarily represent the intentions of my employer.
Regarding you speculations over the future of the joint middleware platform I would suggest you read carefully the following minisite:
Especially the FAQ under:
I think you will find there the direction.
I would suggest to bare in mind that Fusion Applications development has already started and it is heavily based on Oracle ADF and the SOA Suite...


Avi Rosenthal said...

dana, Thanks for the comment clarifying your view.

ronen, thanks for the refernces which I read carefully.
These refernces are marketing publications, However the following messages can be extracted from them:
1. The two SOA product lines Oracle and BEA will be supported and evolved.The customers wil have a choice between them.
2. For the Long Term the brnad name will be Fusion Middleware
3. If I understood it correctly, Oracle hinting that BEA WebLogic will be the prefered Application Server for the Long Term.
The long term content of the branded Fusion Middleware is not defined: Will it be based mostly on Oracle's SOA products or on BEA's suite or on both? probably it will include part of both. Which product line wil dominate? The answer could be speculated. After reading the refernces I still support my prediction that BEA products will dominate. In addition to the Software AG - WebMethods example, after Sun's SeeBeyond aquistion the aquired company's SOA suite is the core of Sun's SOA solution.

Ronen said...


Well, I beg to differ on your predictions regarding the SOA platform, given the larger install base and the higher level of integration between the various components especially in the 11g SOA Suite which is built with a SCA infrastructure. On the other hand the new SCA infrastructure allows to integrate various integration components more easily so it might be possible to cross integrate different components for example BPEL Process Manager with ALESB. I am pretty sure though that the infrastructure will remain the same as BEA doesn't even have its own SCA infrastructure but is using some open source one, the Codehaus if I'm not mistaken.

This is all speculation of course as unfortunately I have no insiders knowledge and if I did, I wouldn't write it here ;-)

About Me

Share on Facebook